Differential Evolution Enhanced with Composite Population Information

Jingliang Liao,Yiqiao Cai,Yonghong Chen,Tian Wang,Hui Tian

2015-01-01

Journal of Digital Information Management

Abstract:Differential evolution (DE) is a simple and powerful evolutionary algorithm, which has been successfully used in various scientific and engineering fields. Generally, the base and difference vectors of the mutation operator in most of DE are randomly selected from the current population. Additionally, the population information is not fully exploited in the design of DE. In order to alleviate these drawbacks and enhance the performance of DE, this study presents a DE framework with Composite Population Information based mutation operator (DE-CPI) for global numerical optimization. In DE-CPI, the ring topology is employed to define a neighborhood for each individual and then the direction information with the neighbors is introduced into the mutation operator of DE. By this way, the composite population information, i.e., neighborhood and direction information, can be fully and simultaneously utilized in DE-CPI to guide the search of DE. In order to evaluate the effectiveness of the proposed method, DE-CPI is incorporated into the original DE algorithms, as well as several advanced DE variants. Experimental results clearly show that DE-CPI is able to enhance the performance of most of the DE algorithms studied. Subject Categories and Descriptors I.5.3 [Clustering]; Algorithms G.1.6 [Optimization] General Terms: Evolutionary Algorithms, Neighborhood